Steel surface defects recognition based on multi-label classifier with hyper-sphere support vector machine

Aiming at steel surface defects recognition, this paper proposed a multi-label classifier with hyper-sphere support vector machine (HSVM-MC). Firstly, in order to describe steel surface defects well, a new set of edge distance statistical features are proposed. Then, based on twin-hypersphere support vector machine (THSVM), HSVM-MC is built by introducing binary relevance(BR) idea. This novel classifier has merits of THSVM and BR. Moreover, it can make up the shortcomings of BR method. Experimental resuslts show that the novel multi-label classificatin method has superior accuracy and efficiency for steel surface defects recognition.